Re: [PATCH 5/6] testsuite: Introduce $inferior_spawn_id

[Pedro Alves <palves@redhat.com>]

On 04/16/2015 06:14 PM, Pedro Alves wrote:
> On 04/16/2015 05:55 PM, Antoine Tremblay wrote:
>> I have a question regarding noinferiorio and it's future usage...
>>
>> As the condition with noinferiorio in interrupt.exp is now  :
>>
>> if {[target_info exists gdb,noinferiorio] && $inferior_spawn_id == 
>> $gdb_spawn_id}
>>
>> noinferiorio is effectively bypassed when we are using gdbserver, even 
>> as noinferirorio is true.
>>
>> But what if the board or simulator really can't handle io at all, and 
>> that setvbuf would not work.. then there is no option to disable io 
>> tests in that case ?
> 
> Hmm, but why would such a board be using gdbserver-support.exp?  Can you
> expand a little?
> 
>>
>> Is the intention to remove noinferiorio from gdbserver-base.exp and keep 
>> the tests with the if {[target_info exists gdb,noinferiorio]} check ?
> 
> The idea was make all tests that rely on inferior io make use
> of $inferior_spawn_id, like interrupt.exp.  My thought was that once that
> is done, we can either remove noinferiorio from gdbserver-base.exp or
> leave it, it wouldn't matter, as for gdbserver testing, would be
> always $inferior_spawn_id != $gdb_spawn_id.  But it sounds like you have
> some environment that may make that troublesome.  I'll need to know more
> about it though.
> 

FYI, I now finished the transition to inferior_spawn_id in the
whole testsuite, and posted it as a series here:

  https://sourceware.org/ml/gdb-patches/2015-04/msg00776.html

the last patch removes noinferiorio from gdbserver-base.exp.

Thanks,
Pedro Alves